Reviewing Azure Databricks and Data Lake Analytics

Databricks is a recent addition to Azure that is greatly influencing the technology choices that people are making when determining how to process data.  Prior to the introduction of Databricks to Azure in March of 2018, if you had a lot of unstructured data which was stored in HDFS clusters, and wanted to analyze it in a scalable fashion, the choice was Data Lake and using USQL with Data Lake Analytics.  With the introduction of Databricks, there is now a choice for analysis between Data Lake Analytics and Databricks for analyzing data.

Analyzing Data with Data Lake Analytics

Data Lake Analytics offers many of the same features as Databricks.  You can write code to analyze data and the analysis can be automatically parallelized to scale.  Microsoft has released a new version of Data Lake, which they are calling Data Lake Storage Gen2 to improve the performance of analysis performed with Data Lakes.  The difference, between the old version and the new one, is the hierarchical namespace to Azure Blob Storage which provides an indexing capability which means that operations can be performed on a directory rather than enumerating through all of the data.  Data stored within a Data Lake can be accessed just like HDFS and Microsoft has provided a new driver for accessing data in a Data Lake which can be used with SQL Data Warehouse, HDinsight and Databricks.  With Data Lake Analytics, the data analysis is designed to be performed in U-SQL. While it supports R and Python libraries, users of the technology will need to get up to speed on U-SQL which is a lot like C#.  This knowledge needs to be learned. Since U-SQL is so new, only a few years old, there is not a large number of people who are familiar with it.

Analyzing Data with Databricks

When analyzing data with Databricks, there are three different languages which you can use: R, Scala, and Python.  Data can be read in from a variety of different Azure Storage options, including Blob Storage, Data Lake, and by using a JDBC connection. You can also connect to Azure SQL DB, as well as Azure SQL Data Warehouse. Since there are three different languages which can be used, there is no reason to learn a new language as most people are already very familiar with at least one of the three supported languages.

In addition to the ability to develop code, Databricks offers some other features which are not found in Data Lake Analytics.  Many projects anticipate that people are going to be working in teams and will need to have an environment to share code and version it.  This capability is baked into Azure Databricks as it provides an environment for sharing data with others and natively saving the data to a GitHub repository.  The development environment is Jupyter Notebooks which provides a great way to document the code and include data samples, all at the same time.  Databricks also includes a job schedule component so that work created in Databricks can use a native scheduler which has the ability to retry and send configurable messages on error or completion.  These additional features, plus the ability to code in a language which is already widely used in the industry, give Databricks the edge in determining which technology to use going forward.

Yours Always

Ginger Grant

Data aficionado et SQL Raconteur

Preparing for Exam 70-774 – Perform Cloud Data Science with Azure Machine Learning

There are a number of reasons why you might want to take a Microsoft cert exam. Maybe you want to focus your studies on a tangible thing, or you think it will help further your career, or you work for a Microsoft Partner and they required a certain number of people to pass the exam to maintain their current partner status.  I am not going to get into the long argument regarding whether or not a cert will help you in your career, or not, I can tell you why you might want to take the 70-774 exam. Machine Learning, or Data Science if you prefer, is an important analytic skill to have to analyze data.  I believe that it will only become more useful overtime. Azure Machine Learning is a good tool for learning the analysis process.  Once you have the concepts down, then should you need to use other tools to perform analysis it is just a matter of learning a new tool.  I talk to a number of people who are trying to learn new things, and the study them in their spare time.  It’s very easy to spend time vaguely studying something, but you may find that having a target set of items to study will focus your time, and as a bonus you get a neat badge and some measure of proof that you were spending time on the computer learning new things and not just watching cat videos.

Exam 70-774 Preparation Tips


While you could always buy the book for the exam (shameless plug as I was one of the authors), the book will not be enough and you will still need to write some code, and do some additional studying. This exam one of two needed for the MCSA in Data Science and you an take the exams in any order. The best place to start is by first looking at the 70-774 exam reference page from Microsoft.  There are four different sections in the exam, and I have created some links for each section which will help you prepare for the exam. In studying for exams in the past, the best way I have found to prepare is to look at everything on the outline and make sure that I know it.

Prepare Data for Analysis in Azure Machine Learning and Export from Azure Machine Learning

Normalizing Data
https://docs.microsoft.com/en-us/azure/machine-learning/studio-module-reference/normalize-data

TanH
https://reference.wolfram.com/language/ref/Tanh.html

ZScore
http://stattrek.com/statistics/dictionary.aspx?definition=z-score
http://howto.commetrics.com/methodology/statistics/normalization/

Min Max
https://www.quora.com/What-is-the-meaning-of-min-max-normalization

PCA
https://docs.microsoft.com/en-us/azure/machine-learning/studio-module-reference/principal-component-analysis
https://docs.microsoft.com/en-us/azure/machine-learning/studio-module-reference/principal-component-analysis
https://stackoverflow.com/questions/9590114/importance-of-pca-or-svd-in-machine-learning

SVD
http://andrew.gibiansky.com/blog/mathematics/cool-linear-algebra-singular-value-decomposition/

Canonical-correlation analysis (CCA)
https://en.wikipedia.org/wiki/Canonical_correlation

Singular Value Decomposition (SVD)
http://andrew.gibiansky.com/blog/mathematics/cool-linear-algebra-singular-value-decomposition/

Develop Machine Learning Models

Team Data Science
https://docs.microsoft.com/fi-fi/azure/machine-learning/team-data-science-process/python-data-access

K-Means
https://www.datascience.com/blog/k-means-clustering

Confusion Matrix
http://www.dataschool.io/simple-guide-to-confusion-matrix-terminology/
https://en.wikipedia.org/wiki/Confusion_matrix
https://en.wikipedia.org/wiki/F1_score

Ordinal Regression
https://en.wikipedia.org/wiki/Ordinal_regression

Poisson regression
https://en.wikipedia.org/wiki/Poisson_regression

Mean Absolute Error and Root Mean Squared Error
http://www.eumetrain.org/data/4/451/english/msg/ver_cont_var/uos3/uos3_ko1.htm

Cross Validation
https://towardsdatascience.com/cross-validation-in-machine-learning-72924a69872f

Operationalize and Manage Azure Machine Learning Services

Connect to a published Machine Learning web service
https://docs.microsoft.com/en-us/azure/machine-learning/studio/publish-a-machine-learning-web-service
https://docs.microsoft.com/en-us/azure/machine-learning/studio/consume-web-service-with-web-app-template
https://docs.microsoft.com/en-us/azure/machine-learning/studio/manage-new-webservice

Use Other Services for Machine Learning

Microsoft Cognitive Toolkit
https://www.microsoft.com/en-us/cognitive-toolkit/

BrainScript
https://docs.microsoft.com/en-us/cognitive-toolkit/brainscript-basic-concepts

Streamline development by using existing resources
https://docs.microsoft.com/en-us/azure/machine-learning/studio/gallery-how-to-use-contribute-publish
Perform database analytics by using SQL Server R Services on Azure
https://docs.microsoft.com/en-us/azure/machine-learning/data-science-virtual-machine/provision-vm
https://docs.microsoft.com/en-us/machine-learning-server/install/r-server-vm-data-science
https://journal.r-project.org/archive/2009-2/RJournal_2009-2_Williams.pdf
http://blog.revolutionanalytics.com/2017/07/xgboost-support-added-to-rattle.html
https://github.com/JohnLangford/vowpal_wabbit/wiki

I hope you have found this test preparation material helpful.  If you passed the exam, let me know by sending me a comment.

Yours Always,

Ginger Grant

Data aficionado et SQL Raconteur

Azure Machine Learning Workbench

Microsoft released Azure Machine Learning Workbench at the Ignite conference on September 25, 2017 as a public preview.  This tool is a new tool which they are adding to their Azure ecosystem, which includes the machine learning tool they introduced three years ago, Azure Machine Learning Studio. Microsoft has said they plan on keeping both products. When asked about the two products, they said that the earlier tool, Azure Machine Learning Studio, is targeted to developers who wanted to add machine learning to their current applications, as it is an easy to use tool that doesn’t require a person to be a trained data scientist.  Azure Machine Learning Workbench is targeted to data scientists who want to bring in other libraries, like TensorFlow for Python, and delve deep into the data.

Microsoft Moves into Machine Learning Management

Microsoft is looking for Azure Machine Learning Workbench for more than a tool to use for Machine Learning analysis. It is part of a system to manage and monitor the deployment of machine learning solutions with Azure Machine Learning Model Management. The management aspects are part of the application installation.  To install the Azure Machine Learning Workbench, the application download is available only by creating an account in Microsoft’s Azure environment, where a Machine Learning Model Management resource will be created as part of the install. Within this resource, you will be directed to create a virtual environment in Azure where you will be deploying and managing Machine Learning models.

This migration into management of machine learning components is part of a pattern first seen on the on-premises version of data science functionality.  First Microsoft helped companies manage the deployment of R code with SQL Server 2016 which includes the ability to move R code into SQL Server.  Providing this capability decreased the time it took to implement a data science solution by providing a means for the code can be deployed easily without the need for the R code to be re-written or included in another application. SQL Server 2017 expanded on this idea by allowing Python code to be deployed into SQL Server as well.  With the cloud service Model Management, Microsoft is hoping to centralize the implementation so that all Machine Learning services created can be managed in one place.

Hybrid Cloud, Desktop, and Python

While you must have an Azure account to use the Machine Learning Workbench, the application is designed to run on a locally on either a Mac or Windows computer.  There is a developer edition of the tool so that one can learn the tool and not incur a bill, which is the case with the previous product, Azure Machine Learning.  The download of Machine Learning Workbench must be accessed within an Azure account and is installed to your local computer.  When running the application from your computer, the application will prompt to log into your Azure account to load Azure Machine Learning Workbench.

The application is designed to use and create Python code.  Azure Machine Learning Workbench does not contain any accommodation to incorporate machine learning components written in R, just Python.  If you have created machine learning components using R, they can be incorporated into the Azure Machine Learning Model Management if you create webservices which encapsulate the R code. The R code does not interface into Workbench, but can be made to be a part of the managed projectes in Azure. While it is possible to create a webservice for R with the earlier product with Azure Machine Learning, there is no direct way to include R with Azure Machine Learning Workbench.  There are a number of sample templates to get started using Python templates including the ubiquitous Iris dataset, Linear regression and several others.   Once the project is created, you can use your favorite IDE, it creates python code which can be read anywhere.

Staying within Machine Learning Workbench application allows you access to arguably one of the neatest parts of the Machine Learning Workbench, the data parser. This tool which was originally code-named project Pendleton and designed to be an intuitive way to modify the contents of data even better than the previous leader in parsing data, Power BI’s Power Query.

You can select the option “Derive column by example” or “Split Column by Example” and then start typing in a new column.  For example, if you want to separate a column which contains the date and the time, if you right click on that date column and select “Split Column by Example” then type the date in the new column provided, the application will immediately determine that you want two columns and crate them. The date column and a time column be created for you after typing in one date.  After the sample columns have been created, you can approve the change or reject it if does not work how you want to.

Like Power Query, each change made to the data is included in the window called Steps on the right side of the application window. When you are done modifying the data, right click on the Data Preparations source icon, which in my example is called UFO Clean, to and the UI changes made to the data are used to create Python code to perform the changes. The generated Python code can be used to the source data programmatically.

The next step in the process is to write the python code needed to evaluate the data and create a model which would in my case determine where and when you are most likely to see the next UFO based on the dataset I have included in my project.  Unlike it’s counterpart Azure Machine Learning, you will need to know how to write the necessary code needed to create a machine learning analysis in Python for Azure Machine Learning Workbench. One could write the Python code to create a machine learning analysis in any Python editor.  If you chose to use Azure Machine Learning, the Python library scikit-learn is installed as part of the application.  Other libraries which you may want to use, such as the common library matplot, you will need to load within Azure Machine Learning Workbench.

Web Service: How Azure Machine Learning Workbench Solutions are Deployed

To deploy a package, you will need to export the completed model serialized Python object, with the Python Module, Pickle. This will create a file with the suffix of pkl, which is the file that you will be deploying. Azure Machine Learning Workbench expects that you will be deploying via Docker containers or creating an Azure cluster.  You will need to register the Docker container in the Machine Learning Container for it to be deployed.

Yours Always,

Ginger Grant

Data aficionado et SQL Raconteur

Running Power BI Locally with the Power BI Report Server

Power BI Now Available on your Local Server

Power BI: Now available without being on the cloud

Microsoft had an lot of announcements about Power BI this week, so many that it was easy to miss some of the finer details, including those which are going to be important in making decisions going forward.  Since the announcements are changes which will be effective soon, in the case of the free tier of Power BI on June 1, and released “… generally available late in the second quarter of 2017” this will give Power BI users time to adjust to the changes. In a nutshell, Microsoft has announced they are adding a cloud service called Power BI Premium which will allow people to create capacity instead of per-user licenses, the free edition will no longer to be able to share files, Power BI Embedded is going to be migrated to the Power BI Service from Azure, and finally, at long last, it will be possible to run Power BI reports locally and without needing anything in the cloud.

Running Power BI without a Cloud

It is not possible to run Power BI reports locally right now, but sometime before the 1st of July 2017,  users who have SQL Server 2016 Enterprise Edition per-core and active Software Assurance [SA] can deploy Power BI Report Server.  This means that no one is going to have to wait for SQL Server 2017 for Power BI on premise as it will be available sometime in June.  The functionality in SQL Server 2017 SQL Server Reporting Server [SSRS]. Community Technology Preview edition is going to be available in Power BI Report Server, with the addition of the ability to include custom visuals and many data sources, which the CTP version did not do. The Power BI Server includes all of the functionality of SSRS This means that users will not need an SSRS Server and a Power BI Server, as the Power BI Server will be able to do both.  If you want to migrate all of the reports created in SSRS from 2008 R2, and SSRS Mobile Reports, you can migrate these reports to the new Power BI Report Server. You can use Power BI Reporting Server for reports created on earlier versions, as long as you have a version of SQL Server 2016 Enterprise per-core edition with SA. The Power BI Report Server will be a separate install with separate release schedules, which currently are planned about once a quarter. Power BI Report Server will also be able to publish reports to mobile devices as well. If the reports uses data in the cloud, you can employ a Data Gateway as the Power BI Reporting Server can use the gateway to access cloud data. Of course if all of the data in the report is located on-premises, no gateway will be required.

Power BI Pro Licenses for On-Premise Reporting

While there is going to be no additional cost for running reports locally, or looking at them, creating and sharing reports for the Power BI Report will require a Power BI Pro License.  The Power BI Desktop is going to be free, and there is still going to be a free version of Power BI. There will also be a  new desktop version of Power BI for Reporting Services which will be on the same version as the Server, which will have fewer updates. This means if you support Power BI Service Reports and Power BI Report Server Reports you will have two versions of the Desktop, the Reporting Services Power BI Desktop and the Power BI Service Desktop.  Both are designed to run on the same machine. So far I have not had any problems having both other than remembering which is which as the icons are the same.  You have to load the software to see that the top line has (Report Server).

Starting June 1, free Power BI license holders will no longer be able to share reports.  Reports created with a free license can be viewed only by the person with the free account.

Power BI Desktop does not have Dashboards, and neither will Power BI

When it is released, Power BI Report Server will be displaying reports created from the Power BI Desktop.  Dashboards are not created in the Power BI Desktop application, meaning that there will be no Power BI Dashboards in the Power BI Report Server.  While this may change in a later release, it is not available in the first release, which also does not support R or custom visuals either.  To display and distribute dashboards, use the Power BI service.

I am sure there will be more announcements about this and other upcoming Power BI features. Many will most likely be announced at Microsoft’s Data Summit Conference in June, which I will fortunately have the opportunity to attend.  If you are going to be there as well, drop me a line or ping me on twitter at @desertislesql and perhaps we can meet in person.
 ***Update I have a post which covers the released version of Power BI Report Server.  Click here to find what was changed since this post was written.
Yours Always

Ginger Grant

Data aficionado et SQL Raconteur

 

 

Security Updates to Power BI

Office 365 Admin Screen for granting Power BI Admin rights

Office 365 Admin Screen for granting Power BI Admin rights

In the past month, Microsoft has made a number of security changes to Power BI. The first one, is not really a feature update, but a PowerShell replacement. No longer do you need to use PowerShell to become a Power BI Admin. Any Office 365 Admin can grant Power BI Admin permissions via this screen in the Admin Center. The Power BI Admin role was first created in October, but the screen was not complete, which was just fixed in February.

Power BI Security Changed from Tenant Only

People who have been granted Power BI administrator rights will also notice a modification to the Admin screen. The March 2017 update to Power BI provides a major change to the security model in Power BI. Previously all the security settings were set at the Tenant Level, meaning that all the privileges were granted to all users. If I wanted to allow one group within the organization to be able to publish reports to the web, but I did not want to allow everyone to publish reports to the web, there was no way that this could be accomplished. All that has changed. It is now possible to include or exclude groups of users from having rights in Power BI. Users can be classified into security groups in Azure Active Directory, either through the Office 365 Admin Center or via the Azure AD Admin Center. Once created the security groups can be used in Power BI. Security Groups are not the same thing as the groups created in Power BI when a new work group is created.

Using Security Groups in Power BI Admin

PowerBINonTenantAdmin

Power BI Admin Portal

The new Power BI Screen looks different. It now lists which rights can be specified to different groups of users. Share content to external users, Export Data, Export reports as PowerPoint presentations, Printing dashboards and reports, Content pack publishing, and Use Analyze in Excel with on-premises datasets now have the ability to be assigned to security groups so that the rights do not have to be the same throughout the entire tenant.

Unfortunately, some of the permissions are still tenant based. For example, the setting Publish to web, which is one permission I would definitely like to turn on only for some users, is still only available as a tenant level option.  These security changes are a welcome improvement to the product as they provide more options for administrators to grant rights to Power BI.

 

Yours Always

Ginger Grant

Data aficionado et SQL Raconteur

 

PolyBase – Another Method for Creating a Stretch Database in SQL Server 2016

PolyBase, which was released with SQL Server 2016, provides another method to access live data either locally or in the cloud, very similar to the SQL Server Stretch database feature. Polybase can also provide the ability to provide a more cost-effective availability for cold data, streamlines on-premises data maintenance, and keeps data secure even during migration. Polybase differs from Stretch database in a few ways, as the SQL must be different, the speed is noticeably slower, and it is a lot less expensive. The cost is significantly less because storing data in a Azure blob store starts at 1 cent a month and Stretch database starts at $2.50 an hour. In this post,I will show how to take data which was archived due to the age of the data, which was created in 2012 and store it in an Azure Blob Storage file which will be available via Polybase when I needed.

Implementing a PolyBase Stretch Database

PolyBase is an optional feature of SQL Server 2016, and the Instance Feature PolyBase Query Service for External Data needs to be selected as part of the installation process. Two services are ssmsexternalinstalled with the feature, SQL Server PolyBase Engine and SQL Server PolyBase Data Movement. Both of these services must be running and TCP/IP must be enabled for PolyBase to work. Either check for those services or run the query SELECT SERVERPROPERTY (‘IsPolybaseInstalled’) AS IsPolybaseInstalled; which will return a 1 when PolyBase has been installed. The next step is to tell SQL Server what the external source is by configuring which Hadoop Connectivity will be used.


EXEC sp_configure ‘hadoop connectivity’, 7;
GO
RECONFIGURE;

To use Azure Blob Storage, there are 3 different options, 1,4 and 7, which include the ability to access blob storage. These options also allow for various other Hadoop engines to be accessed as well. I have selected 7, which also allows me to use Hortonworks on Linux. For more information on the connectivity options, check out this link. A secure database ke is required to make the connection a well. The next step is to let SQL Server know where exactly the blob storage data is stored, which creates an entry in the External Data Sources. PolyBase needs to know how the data is formatted, which will be stored in an External File Format. By definition, Hadoop data has no schema, so a schema is going to have to be created and stored in an external table. The location for each of these items is stored within SQL Server Management studio as shown here.

This code will create a secure key using a password I made up

CREATE MASTER KEY ENCRYPTION BY PASSWORD = 'P@ssW0rdPolybase'

A PolyBase will need the Azure Blob Storage key to be able to make a connection.

CREATE DATABASE SCOPED CREDENTIAL AzureStorageCred
WITH IDENTITY = 'user', Secret = 'lEGL66LiK2KE2U0WEb435PH15BwFOInrHqQ1AJigjoRVfEOrOge+TLbBNu861cqbC+a26io92o5sw/b+OTD/C4w=='; --Note: Replace with your Blob Storage key


CREATE EXTERNAL DATA SOURCE AzureBlobStorage with (
TYPE = HADOOP,
LOCATION ='wasbs://gingeriottest@gingertestiotblobstore.blob.core.windows.net', --This is the location of the folder inside blob storage containing my data
CREDENTIAL = AzureStorageCred
);

The data is stored in blob storage as a CSV, so I will need to create that file Format

CREATE EXTERNAL FILE FORMAT CSV WITH (
FORMAT_TYPE = DELIMITEDTEXT
,FORMAT_OPTIONS (FIELD_TERMINATOR =',', DATE_FORMAT='MM/dd/yyyy') )

Lastly, a table definition must be created for the file so that it can be accessed as a table

CREATE External TABLE OntimePerformance2012sampleall(
[DimAirlineKey] int,
[DimOriginAirportKey] int,
[DimArrivalAirportKey] int,
[DimCancellationReasonKey] int,
[DimDelayLengthKey] int,
[DimDepartureBlockKey] int,
[DimArrivalBlockKey] int,
[DimDistanceGroupKey] int,
[FlightDateKey] int,
[FlightNumber] varchar(50),
[ScheduleDepartureTime] int,
[ActualDepartureTime] int,
[DepartureDelayInMinutes] int,
[TaxiOutTime] int,
[TaxiInTime] int,
[ScheduleArrivalTime] int,
[ActualArrivalTime] int,
[ArrivalDelayInMinutes] int,
[ScheduleElapsedTime] int,
[ActualElapsedTime] int ,
[DistanceInMiles] int,
[CarrierDelayInMinutes] int,
[WeatherDelayInMinutes] int,
[NASDelayInMinutes] int,
[SecurityDelayInMinutes] int
)

WITH
(   LOCATION = '/OntimePerformance2012.csv',
DATA_SOURCE = AzureBlobStorage,
FILE_FORMAT = CSV
)

 

All of the external tables created can be found in the DMV sys.external_tables or by looking at sys.tables where the new column is_external is set to 1.  Once the setup of PolyBase is complete, I can query the external and internal tables all at the same time, like I do here.

 

SELECT count(*)
, r.cancellationReason
, d.CalendarYear
FROM dbo.CancellationReason r
JOIN dbo.OntimePerformance2012sampleall c
ON c.dimcancellationReasonKey = r.CancellationReasonKey
JOIN dbo.DimDate d
ON c.flightDateKey = d.DateKey
GROUP BY r.cancellationReason, d.CalendarYear

PolyBase Table Considerations

It is not possible to edit the data in PolyBase, as the data is stored read only. Also since the data is being pointed to externally, it cannot be indexed and querying it is not terribly quick. All of the data will be brought down and then filtered, unless PolyBase is configured to improve performance, which I will discuss in a future blog post.

PolyBase and Azure Blob Storage provides the ability to access data stored offsite in a secure location in a cost effective manner. If you saw my presentation on this topic at PASS Summit or have questions how to make it work in your environment, please contact me on twitter or write a comment on my blog.

Yours Always

Ginger Grant

Data aficionado et SQL Raconteur

Analyzing JSON in U-SQL

In USQL there are built-in extractors for parsing text, comma delimited or tab delimined files. Once again, parsing JSON becomes problematic. There is a solution built into USQL, write some C# code to extend it or use someone else’s C# code to extend USQL. Since I wanted to parse JSON, fortunately there are libraries available on github containing the information required to do it. Download the github package and open up the Microsoft.Analytics.Samples project in Visual Studio. When I did this the first time, there was a problem loading the Newtonsoft.Json reference, so I right clicked on the references and downloaded the missing parts again. Build the solution and check out the code in the directory …Examples\DataFormats\Microsoft.Analytics.Samples.Formats\bin\Debug\ . There will be two DLLs, Microsoft.Analytics.Samples.Formats.dll and Newtonsoft.Json.dll. These dlls then need to be registered in Data Lake Analytics and locally if you chose to run your USQL locally. As at some point the goal is to run from within Data Lake analytics, you will need to copy both of these dlls to the data lake. I created a folder for the dlls called Assemblies, and ran this command


USE DATABASE [master];
CREATE ASSEMBLY [Newtonsoft.Json] FROM @"/Assemblies/Newtonsoft.Json.dll";
CREATE ASSEMBLY [Microsoft.Analytics.Samples.Formats] FROM @"Assemblies/Microsoft.Analytics.Samples.Formats.dll";

Notice I told the USQL where to find the dlls, in the Assemblies folder. This step only needs to be completed once per data lake. After this job successfully runs, then the dlls which allow the JSON to be parsed, can be referenced.

Here is my sample JSON, which I have copied to the folder Samples/Data/TestNew.Json, in the Data Lake

{
"appInstanceId": "357ced1e-cf05-459c-9317-794bq24f61c2",
"firmwareVersion": "1.0.2.4",
"serialNumber": "254542-694967",
"Side": "0",
"Latitude": "33.8848744",
"Longitude": "-128.403276",
"GeneratedDate": "2016-10-04T21:18:19Z"
}

Now that I have added the JSON to the Data Lake and the assemblies have been added, I can write some USQL to Parse the JSON. First I will need to reference the libraries, then create a schema, as there is no schema for a Data Lake. After those steps are completed, it’s possible to write SQL to query a JSON file. There is no UI to look at the results, so the results will be writing to a file. I am going to output the data to a csv file called JSONOutput.csv. Here’s the code to do that.

REFERENCE ASSEMBLY [Newtonsoft.Json];
REFERENCE ASSEMBLY [Microsoft.Analytics.Samples.Formats];

DECLARE @infile string="/Samples/Data/TestNew.json";

@logSchema =
EXTRACT name string
, appInstanceId string
, firmwareVersion string
, serialNumber string
, Side string
, Latitude float
, Longitude float
FROM @infile
USING new Microsoft.Analytics.Samples.Formats.Json.JsonExtractor();

@testthis = SELECT appInstanceId
, COUNT(*) AS LocationCount
FROM @logSchema
GROUP BY appInstanceId;

OUTPUT @testthis
TO "/Samples/Data/JSONoutput.csv"
USING Outputters.Csv();

vsrunjson

Using Visual Studio, I am running the USQL Job. There isn’t much data to parse, and you can see in the summary widows that it took 21 seconds to prepare, and 33 seconds to run.

When go to the web and look at the Data Lake Analytics page, I can also see that the job completed. I have noticed that this appears pretty close to the same time on the web and on visual studio.

azuredlscreen

Clicking on the bar graph represented by today will allow me to select the job which ran, showing the same screen as appears in Visual Studio.

Thanks to Erik Zwiefel and Mark Vaillancourt b | t both of Microsoft for helping me figure out the process to use JSON in Data Lake Analytics, as I didn’t understand the steps which are required to parse JSON. I hope this blog makes it possible for you to figure out how to make it work.

 

 

Yours Always

Ginger Grant

Data aficionado et SQL Raconteur

 

 

Using Visual Studio for U-SQL Data Analytics Jobs

The pricing for USQL is based upon how many Analytic Units and Completed Jobs. To decrease the amount of money being spent, it would be most efficient if only completed jobs ran on SQL, not the 27 times the job was run to debug it. Fortunately, all of the debugging can be performed locally and only working jobs need to be run. Another thing that you may notice if you are exclusively using the Azure Portal for running Data Lake Analytic jobs is there is no way to actually save a job. Once the job is completed, you can review the job then click on the View Script button. Don’t rely on the button though, because for reasons unknown, sometimes the View Script Button is not enabled, meaning that it is not possible to see what ran.

Data Lake Analytics Setup for Visual Studio

There are a few steps required before any code is run. If the Data Lake Analytics Tools are not installed within Visual Studio, download them here and install them. When the tools are installed, the menu item Data Lake appears in Visual studio. The second step is to model your PC with the same file structure as your data lake. The default location which the Data Lake tools will look for your data structure is C:\Users\<<insertyourname>>\AppData\Local\USQLDataRoot . What this means is if you have folders and subfolders created in your data lake, your PC needs to have the same structure, including the data.

Running Data Lake Jobs Locally

If you take a look at the screen picture of Visual Studio with the data lake installed, you will notice a series of buttons at the top of the screen. The middle button currently is set to (Local). The drop down box at the top of the top of the screen will allow you to set the job to either your Azure Data Lake Analytics instance or locally. If it is running locally there will be no charges incurred on Azure. Also in Visual Studio, of course you can save the name of the USQL file.

vsdatalakescript

 

When the context is switched to the Data Lake Analytics instance from Azure, you will probably want to check out the Summit button. There is only one option, Advanced. In this window, you can change the Job name. It is default set to the name of the script being run, but if you are running the same script over and over, you may wish to change this name so that the different instances can be identified. Parallelism can also be set to the value that is actually being used in the job. Take a look at the job view, which is the tab to the left of Script. This screen shows the processes in use when the job is run and set the value accordingly. You will be charged for the Parallelism value that is set, not the amount actually used. Setting to a lower value can decrease the cost of running a job.

The tab on the far left, shows the job with the same view shown in the Azure Portal Stream Analytics job. That screen is shown below.

usqlscreenRunning on Visual Studio also has the benefit of less changing between screens than the Azure Portal, which is another reason to develop here. Now that I have this environment set up, I plan on writing all of my Data Lake Analytics jobs here, as I find the development environment works better for me. Let me know what you think of it by commenting below. If you are interested in finding out more about running Data Lake Analytics Jobs, especially if you are trying to parse JSON, please subscribe to my blog as that topic will be in an upcoming post.

Yours Always

Ginger Grant

Data aficionado et SQL Raconteur

Event Hub Troubleshooting

When creating an Azure Event Hub, chances are there will be no errors. This is not always a good thing, as it may mean that the errors exist but do not appear. Maybe the event hub is sending data, but the data cannot be read by a stream analytics job. Maybe the event hub really is working, but nothing appears in the dashboard. If any of these problems sound familiar, this post should help.

Testing the Event Hub

If you don’t have a source of data, like a raspberry pi or a sensor sending data, you can use this guide to create a C# program to send data to your event hub. Chances are though, this code is going to have to be modified even more than the instructions indicate, because the data sent is not in JSON. While it is not a requirement that data sent to the event hub be in JSON, if you want to read it with stream analytics that is one of the acceptable formats needed. If you are using the code provided and you want to insert a record into a database field input01, the message needs to be changed to the following to add the double-quotes and brackets required by JSON.
var message = "{'input01':\"" + Guid.NewGuid().ToString()   + "\"}";

Validating the Event Hubs Receive Messages

To ensure that the event hub is actually receiving data, validation can only be done in the old Azure portal. The service bus icon is two down from the HDInsight elephant. Double-clicking on the service bus namespace will bring up the a list of event hubs. Double-clicking on that will show this screen. This screen picture was taken roughly at 7:10. How many messages are there at 7:00? None.

eventhubbefore

This screen print was taken at 7:17. Notice anything different about the message count at 7:00?

eventhubafter

Oh look, there are 144 Messages which came in between 7:00 and 7:05. This means that everything really was working, I just needed to wait to see them appear. The wait time tends to vary from 10-20 minutes. Perhaps nothing is wrong. Lucky if this is you as you can stop reading

Stream Analytics with Event Hubs

If you are using an event hub to pass data to a stream analytics job, step one, make sure the stream analytics job is started. Created does not mean started as it should say Running as shown in the clip below.

The input for this stream is set to an event hub which has a standard subscription. The basic subscription, which is of course cheaper, has one default consumer group. With a standard subscription multiple consumer groups can be created and more importantly named. When setting up the inputs there is a blank for the name of the consumer group. If you have a basic subscription this will be empty. If it is empty, then the event hub won’t pass data to the stream analytics job. Perhaps there is a way to get a basic event hub to work with a stream analytics job, but I couldn’t make it happen. When I created an event hub with a standard subscription and created a consumer group and added that name to the input of a streaming analytics job, it worked.

If you have found these troubleshooting tips helpful, please subscribe to my blog, as I will be passing along more tips in my next post which will detail the steps of how to get data from the event hub to a Azure Database.

 

Yours Always

Ginger Grant

Data aficionado et SQL Raconteur

 

 

U-SQL and Azure Data Lake Analytics

There are a number of different SQL Flavors–HQL, PL/SQL, MySQL, U-SQL, T-SQL — all of which are a derivative of Ansi-SQL, which is I suppose in today’s parlance, A-SQL. Many people have not heard of U-SQL, which Microsoft introduced on September 28, 2015. Since the announcement was in the Visual Studio Blog, a number of data people may have missed it. U-SQL is meant to combine the ease SQL with the functionality of C# to create a language which can process any kind of data, like videos or text, by creating the ability to customize the code and infinitely scale. This is very useful if for example all of the data is stored in an Azure Data Lake.

Using U-SQL in Azure Data Lake Analytics

In my previous series on Stream Analytics, I wrote some U-SQL. That U-SQL didn’t look much different than Ansi-SQL, which is sort of the point of porting the functionality to a different yet familiar language. Another application which heavily uses U-SQL is Azure Data Lake. Data Lake stores its data in HDInsight, but you don’t need to write hive to query the data, as U-SQL will do it. Like Hive, U-SQL can be used to create a schema on top of some data, and then query it.

For example, to write a query on this csv file stored in a Data Lake, I would need to create the data definition for the data, then I could easily write a statement to query it.

PopsicleDataLake

@searchlog =
EXTRACT SaleDate string,
SaleLocation string,
Lemon   int,
Orange     int,
Temperature     int,
Leaflets int,
Price                      string
FROM "Samples/Data/Popsicle.tsv"
USING Extractors.Tsv();


@testthis = SELECT SaleLocation
, COUNT(*) AS LocationCount
FROM @searchlog
GROUP BY SaleLocation;


OUTPUT @testthis
TO "Samples/Data/Output/SaleLocCount.csv"
USING Outputters.Csv();

In this U-SQL code, I am creating a structure for the data, querying some fields, and writing the output to another file. Make sure that you don’t forget the semi-colons as that will cause errors. Also if any of your fields are blank you will have to code for that as well. From with Data Lake Analytics, the U-SQL is run as a job, creating a new file. Note the time that it took to finish the job.

USQLJob

 

The reason data is stored in a Data Lake is to provide a single storage location for the data, which will be used in analytics. U-SQL provides a powerful tool for getting the data out.

Yours Always

Ginger Grant

Data aficionado et SQL Raconteur